Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: difference between -test- and -contrast- in Statat 12


From   Ricardo Ovaldia <[email protected]>
To   "[email protected]" <[email protected]>
Subject   st: difference between -test- and -contrast- in Statat 12
Date   Thu, 25 Aug 2011 19:15:13 -0700 (PDT)

 
I am using the new -contrast- command in Stata 12 to test a contrast after ANOVA and I get different p-values than when using -test- although the marginal means are the same.
 
I illustrate with the auto data by first making the rep78 variable a 3 level variable and then running an ANOVA follow by -contrast-:

. sysuse auto,clear
. replace rep78=3 if rep78<3
. anova price rep78##foreign                                          
 <output omitted>
. contrast r.foreign@rep78
 
Contrasts of marginal linear predictions
 
Margins      : asbalanced
 
-------------------------------------------------
              |         df           F        P>F
--------------+----------------------------------
foreign@rep78 |
  (1 vs 0) 3  |          1        0.73     0.3965
  (1 vs 0) 4  |          1        0.07     0.7881
  (1 vs 0) 5  |          1        0.80     0.3743
       Joint  |          3        0.53     0.6606
              |
     Residual |         63
-------------------------------------------------
 
---------------------------------------------------------------
              |   Contrast   Std. Err.     [95% Conf. Interval]
--------------+------------------------------------------------
foreign@rep78 |
  (1 vs 0) 3  |  -1529.739   1792.049     -5110.862    2051.385
  (1 vs 0) 4  |   379.8889   1407.262       -2432.3    3192.078
  (1 vs 0) 5  |   2088.167   2333.681     -2575.322    6751.655
---------------------------------------------------------------
 


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index